The use of stock-based pay for sorting: an empirical analysis of compensation for new CEOs

Abstract

Examining stock-based compensation for newly hired CEOs, this article finds that the sensitivity of stock-based pay to performance is higher for new economy, young and volatile firms. Of the components of stock-based pay, it is option grants that generate such variation across firms. It also finds that this cross-firm variation in pay-performance sensitivity is more pronounced for the CEO's first year in office. These findings support the view that firms use stock-based pay to new CEOs for sorting.

Acknowledgement

I thank Chul Chung and an anonymous referee for helpful comments.

Notes

¹ While both the provision of incentives and sorting envision that stock-based pay is used by the firm to motivate or discipline its executives, Bebchuk et al. (2002) argue that stock-based pay is designed by the executives, not by the firm, as camouflage to conceal high levels of pay.

² While most of the papers on this topic are found in the economics and finance literature, there is an excellent survey by Tosi et al. (2000) in the management literature.

³ In the context of compensation for rank-and-file workers, Booth and Frank (1999) and Lazear (2000) find evidence that the use of stock-based pay is related to sorting. See also Oyer and Schaefer (2005) and Nagaoka (2005) for the adoption of broad-based equity grants for sorting purposes.

⁴ As this sensitivity is based on the percent change in stock price, it is more precisely pay-performance semi-elasticity.

⁵ CEOs are usually undiversified and their trading of the company stock and options is implicitly and/or explicitly restricted. The Black–Scholes formula that is based on tradable options may then over-estimate the true value of options to CEOs. In response to this problem, Hall and Murphy (2002), for instance, suggest an alternative method to value executive options that takes CEOs’ risk-averseness and undiversification into account. However, this method requires information about the recipient's utility function, risk-averseness and wealth. As this information is hardly obtainable, the analysis of this article uses the Black–Scholes formula, which is widely used in the literature on CEO compensation including Jensen and Murphy (1990), Yermack (1995), Hall and Liebman (1998), Aggarwal and Samwick (1999), Core and Guay (1999, 2002) and Coles et al. (2006).

⁶ See Prendergast (1999) for a survey of theories relating the use of stock-based pay to the provision of incentives, Oyer (2004) for a theoretical model explaining the use of stock-based pay for retention and Bebchuk et al. (2002) for an argument of managerial rent seeking.

⁷ Main identifiable information is obtained from the data item ‘becamece’, which indicates the calendar date of hiring. An alternative piece of information is from the data item ‘ceoann’, which indicates whether the corresponding individual was the CEO for all or most of the indicated fiscal year. In the data item ‘ceoann’, some CEOs may not be identified as CEO in the year when they were hired, if their predecessor took the position for most of that year. Thus, technically the fiscal year of hiring identified from ‘becamece’ is the same or one year earlier than the year identified from ‘ceoann’. The final sample consists of CEOs who meet this condition.

⁸ For some CEOs, options are granted several times during their first year in office. Each grant is valued separately and then summed over the year.

⁹ More precisely, new economy firms are referred to as companies with primary SIC 3570, 3571, 3572, 3576, 3577, 3661, 3674, 4812, 4813, 5045, 5961, 7370, 7371, 7372 and 7373.

¹⁰ Holmstrom (1979) shows a trade-off between incentives and risk, indicating a negative association between pay-performance sensitivity and risk measures. Since Jensen and Murphy (1990), numerous papers examine this association including Aggaral and Samwick (1999), Jin (2002) and others. In contrast, Prendergast (2002) argues that the trade-off between incentives and risk is rather tenuous, which is also confirmed by several empirical studies including Foss and Laursen (2005).

¹¹ Another determinant of the provision of incentives is CEOs’ career concerns, which can be measured as for instance the number of years remaining as CEO (Gibbons and Murphy, 1992). Due to the lack of precise information about each CEO's retirement year, the effects of CEO career concerns are left uncontrolled in the models. However, it would be a reasonable presumption that the degree of career concerns is similar across new CEOs because all of them just begin their career as CEO.

¹² See Aggarwal and Samwick (1999) for a detailed discussion about a rationale for the use of dollar returns when measuring compensation risk.

¹³ See Hirshleifer and Suh (1992), Guay (1999), Ryan and Wiggins (2002) and Coles et al. (2006) for the use of options to provide risk-taking incentives.

¹⁴ See also Roosenboom and van der Goot (2006) for such evidence in IPO firms.

¹⁵ See Core and Guay (1999) for the use of equity grants to adjust equity holdings to optimal levels.

¹⁶ The correlation coefficient is 0.24 between new economy and young firms, 0.33 between new economy and volatile firms and 0.41 between young and volatile firms.